Guehi et al. (2025) A revisit of Niger river’s major flood events and the role of convective systems with satellite data and hydrological modeling

Identification

• Journal: Journal of Hydrology Regional Studies
• Year: 2025
• Date: 2025-12-12
• Authors: Fatou Josiane Guehi, Rômulo Augusto Jucá Oliveira, Marielle Gosset, Modeste Kacou, Eric-Pascal Zahiri, Thomas Fiolleau
• DOI: 10.1016/j.ejrh.2025.103025

Research Groups

• Laboratoire des Sciences de la Mati`eres, de l′Environnement et de l′Energie Solaire (LASMES), Universit´e F´elix Houphou¨et-Boigny, Abidjan, Cˆote d’Ivoire
• G´eoscience Environnement Toulouse (GET), IRD, CNRS, CNES, Universit´e de Toulouse, Toulouse, France
• Hydro Matters, Le Faget, France
• Laboratoire d’Etudes en G´eophysique et Oc´eanographie Spatiales (LEGOS), CNRS/CNES/IRD/UPS, Universit´e de Toulouse, Toulouse, France

Short Summary

This study develops and applies a novel method combining satellite data and hydrological modeling to quantify the impact of individual mesoscale convective systems (MCSs) on Niger River floods in Niamey, revealing that a small number of MCSs are responsible for a significant portion of extreme flood events, particularly the record 2020 flood.

Objective

• To analyze the role of individual storms in the genesis of strong rises in discharge that could lead to flooding, proposing an original approach for quantifying the impact of individual Mesoscale Convective Systems (MCSs) on the river's hydrological response.
• To revisit and better understand the genesis of major flood events over the Niger basin, specifically focusing on the role of MCSs in the intensity of the "red flood" and the record flood of 2020 in Niamey.

Study Configuration

• Spatial Scale: Upper and middle Niger River basin down to Niamey, West Africa, with a specific focus on the "red flood genesis area" (right-bank tributaries Gorouol, Sirba, Dargol) covering 130,786 square kilometers.
• Temporal Scale: Analysis of MCSs and rainfall from January 2012 to December 2020, with a detailed focus on the 2020 hydrological year (June 1, 2020, to May 31, 2021) for flood events.

Methodology and Data

• Models used:
  • Hydrological model: MGB (Modelo de Grandes Bacias / Large Basins Model), a large-scale semi-distributed rain-discharge hydrologic-hydrodynamic model.
  • MCS tracking algorithm: TOOCAN (Tracking Of Organized Convection Algorithm using a 3 dimensional segmentatioN), version 08.
• Data sources:
  • Satellite rainfall estimates: Integrated Multi-satellitE Retrievals for Global Precipitation Measurement (IMERG) V06 Final Run product (0.1° grid, 30 minute temporal resolution).
  • Satellite convection tracking: TOOCAN product based on geostationary satellite infrared (IR) data (4 km horizontal, 30 minute temporal resolution).
  • In-situ discharge data: Niger Basin Authority (ABN) for the Niamey station.
  • In-situ rainfall data for validation: AMMA-CATCH (Analyse Multidisciplinaire de la Mousson Africaine - Couplage de l′Atmosph`ere Tropicale et Cycle Hydrologique) observatory rain gauge network in Niamey.
  • Digital Elevation Models (DEMs) for hydrological model discretization.

Main Results

• The MGB hydrological model successfully simulated the 2020 red flood, with high performance (Nash-Sutcliffe Efficiency and Kling–Gupta Efficiency of 0.94, Pearson correlation coefficient of 0.97 for the 2020 hydrological year).
• Mesoscale Convective Systems (MCSs) are the primary drivers of the red flood; removing all MCS-generated rainfall from the model forcing eliminated the red flood entirely, reducing the 2020 peak discharge by 74% (from 2741 cubic meters per second to 713 cubic meters per second) and the mean seasonal flow by 73%.
• A small number of MCSs have a disproportionately large impact: the 10 rainiest MCSs in 2020 contributed 34% of the total annual rainfall and 36% of the rainy season rainfall. These 10 MCSs alone accounted for 51% of the 2020 red flood peak discharge and 43% of the mean seasonal flow.
• The single rainiest MCS (MCS-1) reduced the peak discharge by 17% and the mean seasonal flow by 9%.
• The 10 most impactful MCSs were characterized by long lifetimes (14 to 30 hours), large sizes (many exceeding 100,000 square kilometers), strong convective activity (minimum brightness temperatures ranging from 178 K to 186 K), and relatively slow to moderate propagation velocities (5.5 to 15 meters per second).
• These major MCSs exhibited their maximum convective activity and precipitation rates over the study area, with a lasting effect on discharge for up to several weeks and a delay of a few days between rainfall onset and peak discharge.
• The results are robust to uncertainties in satellite data and hydrological model parameters, with the relative impact of MCSs varying only marginally (approximately 1%).

Contributions

• Introduces an original method that integrates satellite-based MCS tracking (TOOCAN) and rainfall estimation (IMERG) with a distributed hydrological model (MGB) to quantify the specific hydrological impact of individual MCSs.
• Provides a detailed, process-based understanding of the link between individual tropical convective storms and large-scale river flooding, moving beyond statistical correlations.
• Quantifies the disproportionate contribution of a small number of intense MCSs to extreme flood events in the Niger River basin, particularly the record 2020 flood in Niamey.
• Identifies key physical characteristics of flood-generating MCSs (size, intensity, duration, propagation speed) that are crucial for improving flood forecasting and early warning systems.
• Demonstrates the potential for transferring this method to analyze flood risk in other tropical watersheds globally.

Funding

• AFD/IRD CECC project

Citation

@article{Guehi2025revisit,
  author = {Guehi, Fatou Josiane and Oliveira, Rômulo Augusto Jucá and Gosset, Marielle and Kacou, Modeste and Zahiri, Eric-Pascal and Fiolleau, Thomas},
  title = {A revisit of Niger river’s major flood events and the role of convective systems with satellite data and hydrological modeling},
  journal = {Journal of Hydrology Regional Studies},
  year = {2025},
  doi = {10.1016/j.ejrh.2025.103025},
  url = {https://doi.org/10.1016/j.ejrh.2025.103025}
}